• Communications for Statistical Applications and Methods
• /
• v.24 no.1
• /
• pp.81-96
• /
• 2017

Cumulative logit random effects models are typically used to analyze longitudinal ordinal data. The random effects covariance matrix is used in the models to demonstrate both subject-specific and time variations. The covariance matrix may also be homogeneous; however, the structure of the covariance matrix is assumed to be homoscedastic and restricted because the matrix is high-dimensional and should be positive definite. To satisfy these restrictions two Cholesky decomposition methods were proposed in linear (mixed) models for the random effects precision matrix and the random effects covariance matrix, respectively: modified Cholesky and moving average Cholesky decompositions. In this paper, we use these two methods to model the random effects precision matrix and the random effects covariance matrix in cumulative logit random effects models for longitudinal ordinal data. The methods are illustrated by a lung cancer data set.

• Journal of the Korean Data and Information Science Society
• /
• v.17 no.4
• /
• pp.1053-1065
• /
• 2006

This paper proposes a method in one-way layouts for longitudinal data with informative drop-out. When dropouts are informative, that is, correlated with unobserved data and/or the previous observed data, the simple imputation methods such as 'last observation carried forward' (LOCF) methods would arise the bias of the testing models. The maximum likelihood procedure combined with a logit model for the drop-out process is proposed to test treatment effects for one factorial designs and compared with LOCF method in two examples.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.26 no.1
• /
• pp.109-116
• /
• 2001

The detection of corporate failures is a subject that has been particularly amenable to cross-sectional financial ratio analysis. In most of firms, however, the financial data are available over past years. Because of this, a model utilizing these longitudinal data could provide useful information on the prediction of bankruptcy. To correctly reflect the longitudinal and firm-specific data, the generalized linear model with assuming the first order AR(autoregressive) process is proposed. The method is motivated by the clinical research that several characteristics are measured repeatedly from individual over the time. The model is compared with several other predictive models to evaluate the performance. By using the financial data from manufacturing corporations in the Korea Stock Exchange (KSE) list, we will discuss some experiences learned from the procedure of sampling scheme, variable transformation, imputation, variable selection, and model evaluation. Finally, implications of the model with repeated measurement and future direction of research will be discussed.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.32 no.6D
• /
• pp.549-556
• /
• 2012

Longitudinal tunnels are defined as tunnels with length of over 1km. Because of Korea's topographical conditions and as safety measures for linear design, many tunnels are inevitably being constructed in Korea. The number of longitudinal tunnels constructed on expressways amounted to 104 as of the end of 2010 with a total length of 192km. Given the increasing demand for tunnels and the increasing length of tunnels, a safety evaluation of longitudinal tunnels needs to be conducted. As such, this study selected design elements, transportation environment and delineation system as elements to check and tried to determine factors influencing road crashes. For this, tunnels have been classified based on history of crashes; ones with crashes and ones without crashes and statistically meaningful explanatory variables were selected. By using these variables, a logit model was development in order to better grasp the factors that directly and strongly influence crashes. The result, related to crashes as well as the analysis were utility tunnel interior materials of driving lane and passing lane, which are related to driver's visibility, lateral width widening to consolidate space in a tunnel, and annual average daily traffic (AADT) per lane. These results may be used in the future as analysis indicators when drawing up plans to prevent crashes in longitudinal tunnels.

• Health Policy and Management
• /
• v.33 no.3
• /
• pp.295-310
• /
• 2023

Background: The long-term care (LTC) group has higher rates of chronic disease and disability registration compared to the general older people population. There is a need to provide integrated medical services and care for LTC group. Consequently, this study aimed to identify medical usage patterns based on the ratings of LTC and the characteristics of benefits usage in the LTC group. Methods: This study employed the National Health Insurance Service Database to analyze the effects of demographic and LTC-related characteristics on medical usage from 2015 to 2019 using a repeated measures analysis. A longitudinal logit model was applied to binary data, while a linear mixed model was utilized for continuous data. Results: In the case of LTC ratings, a positive correlation was observed with overall medical usage. In terms of LTC benefit usage characteristics, a higher overall level of medical usage was found in the group using home care benefits. Detailed analysis by medical institution classification revealed a maintained correlation between care ratings and the volume of medical usage. However, medical usage by classification varied based on the characteristics of LTC benefit usage. Conclusion: This study identified a complex interaction between LTC characteristics and medical usage. Predicting the requisite medical services based on the LTC rating presented a challenge. Consequently, it becomes essential for the LTC group to continuously monitor medical and care needs, even after admission into the LTC system. To facilitate this, it is crucial to devise an LTC rating system that accurately reflects medical needs and to broaden the implementation of integrated medical-care policies.

• The Korean Journal of Applied Statistics
• /
• v.27 no.6
• /
• pp.923-932
• /
• 2014

The Hurdle model can to analyze zero-inflated count data. This model is a mixed model of the logit model for a binary component and a truncated Poisson model of a truncated count component. We propose a new hurdle model with a general heterogeneous random effects covariance matrix to analyze longitudinal zero-inflated count data using modified Cholesky decomposition. This decomposition factors the random effects covariance matrix into generalized autoregressive parameters and innovation variance. The parameters are modeled using (generalized) linear models and estimated with a Bayesian method. We use these methods to carefully analyze a real dataset.

• The Journal of the Korea Contents Association
• /
• v.14 no.12
• /
• pp.753-763
• /
• 2014

After retirement, retirees are exposed to many changes. But, one of the most influential factor on retirement is Social network. Social network is making various relationships with many people. This consists of functional feature and structure feature. This study to systematically investigate social network changes of pre- and post-retirement by using two features. We utilized 2008~2012 'Longitudinal study of ageing' and selected 1,569 retirees above 45 years old as a final subject. This study used STATA 12.0 program for analysing frequency and descriptive statistic. At first, we analyzed personal characteristics and affecting factor on social network of retirees through Panel logit model and fixed effects regression model. Second, we applied multiple panel logit model and fixed effects model to learn factor affecting employment and social network changes. We found that a number of social activities affects social network in the structure feature and support from sons and daughters also influences social network in the functional feature after retirement.

• 한국사회정책
• /
• v.23 no.1
• /
• pp.95-123
• /
• 2016

As Korean society grows rapidly older, a systematic analysis of the determinants of labor supply behavior among the elderly becomes a prerequisite for designing more effective senior employment policies and income security regime for the elderly. Literatures review shows that a majority of previous researches have been ignoring the problem of "endogenous predictor" especially when it comes to the estimation of the effects of the two key variables, education and non-labor income, on labor supply decisions among older people. They have failed to take into consideration the unobserved heterogeneities which might affect both labor supply decisions of the elderly and their levels of education and non-labor income, which means, according to some econometric literatures, that the estimated coefficients of the two predictors can be inconsistent. The paper tries to redress the endogeneity problem by employing a panel logit model with data from the 1st. to 4th. wave of the KLoSA(Korean Longitudinal Survey of Ageing) to estimate the effects of key predictors on the probability of getting jobs among older people(ages of 60 or older). Both a random effects and a fixed effects model reaffirms that non-labor income has a negative effect on the chances of being employed. And a random effects model shows that the effect of education is also negative, as has frequently been reported by previous studies. That means the effects of education and non-labor income on elderly employment remain negative after the effect of unobserved heterogeneities is controled for and the problem of endogenous predictors is redressed through an appropriate panel data analysis. These findings mean, in turn, that when Korean baby-boomers, who had acquired an unprecedentedly higher level of education and were expected to enjoy ever-larger amount of non-labor income than their preceding generations, retires in near future, their incentives to work will become much weaker and the lack of labor-force and the burden of financing increased public pension expenditure will become more troublesome. The paper concludes with recommending some policy initiatives helpful to solve these expected problems.

• Journal of Digital Convergence
• /
• v.17 no.4
• /
• pp.365-379
• /
• 2019

The purpose of this study is to investigate what factors affect cyber-delinquency after examining the previous research focusing on the general strain theory and the delinquency opportunity theory in the existing studies. And as adolescents move from middle school to high school, this study is intended to analyze what factors affect cyber-delinquency from a longitudinal perspective using KCYPS(Korea Child and Youth Panel Survey) elementary 4th grade fourth and seventh data. The adolescence cyber-delinquency probability of occurrence were analyzed through the panel logit fixed-effect model using STATA. And then the cyber-delinquency frequency of adolescents were analyzed through the panel tobit random-effect model. As a result of analyzing the factors affecting cyber-delinquency frequency, Adult media commitment, computer use time, and cell phone dependency increased cyber-delinquency frequency. On the other hand, among the parenting attitudes, the attitude of supervising attentively and adolescents' age-increasing decreased cyber-delinquency frequency.

• Journal of Families and Better Life
• /
• v.27 no.3
• /
• pp.31-43
• /
• 2009

Using the first wave of KLoSA(Korean Longitudinal Study of Ageing) beta version, this study analyzed factors affecting early retired men's subjective life satisfaction through Binary Logit and Multiple Regression. Total 552 men were selected as a sample. The main results of empirical analysis in this study were as follows: The retirement reason(health-) and monthly household income(+) affected whether they were satisfied with the retirement life or not and subjective life satisfaction over all. Especially, the retirement reason(health-) had a stronger effect on whether early retired min were satisfied with the retirement life or not and their subjective life satisfaction than monthly household income revealed significant variable in previous studies. This result represents that the retiree's life satisfaction analysis model must include retiree's characteristics at the time of retirement as well as retiree's current status characteristics or socio-economic characteristics.